Non-uniform clutter suppression method for airborne radar based on rapid environmental perception

Abstract

This invention relates to a method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception. The method includes: the airborne radar entering the perception phase, where several transmitting array elements transmit orthogonal signals, and simultaneously the airborne radar acquires several first coherent pulses received by several receiving array elements; estimating the number of discrete strong clutter blocks on each range ring of the airborne radar using beam scanning; establishing a sparse recovery model of the echo signal and the radar scattering cross section of the clutter block for each range element; combining the number of discrete strong clutter blocks, solving the sparse recovery model using an orthogonal matched pursuit algorithm to obtain an estimated value of the clutter scattering coefficient for each range ring; the airborne radar entering the detection phase and transmitting a linear frequency modulated signal, while simultaneously receiving several second coherent pulses; constructing a structured clutter covariance matrix for each range element; and using the structured clutter covariance matrix to calculate a space-time two-dimensional adaptive weight for clutter suppression. This method achieves rapid perception of the clutter environment.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing technology, specifically relating to a method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception. Background Technology

[0002] Airborne radar is one of the key weapons on the modern battlefield. Due to its high antenna mounting platform, airborne radar has a greater maximum detection range, effectively detecting low-altitude penetrating targets. Its high mobility and flexibility have made it a primary source of battlefield situational awareness in modern warfare. However, when airborne radar operates in a look-down mode, it faces challenges such as high clutter intensity and clutter spectrum broadening caused by different Doppler frequencies from different directions, severely impacting its detection performance. Furthermore, in actual battlefield environments, clutter environments are often non-uniform, making clutter suppression even more difficult.

[0003] Existing technology 1 proposes a non-uniform clutter suppression method. This method uses only a single echo as a training sample, obtains a high-resolution estimate of the clutter spatiotemporal two-dimensional spectrum through sparse recovery techniques, then uses the estimated clutter spatiotemporal spectrum to estimate the clutter covariance matrix, and performs weighted processing on the estimated clutter covariance matrix. Finally, it calculates the corresponding spatiotemporal adaptive filter coefficients based on the estimated clutter covariance matrix to complete non-uniform clutter suppression and target detection. This method can effectively suppress non-uniform clutter, but its drawback is that dictionary mismatch during the solution of the spatiotemporal two-dimensional spectrum introduces sparse recovery errors, severely affecting the clutter suppression performance of the spatiotemporal two-dimensional adaptive processing.

[0004] Existing technology two proposes a space-time adaptive processing method for non-uniform clutter suppression. This method first sets the airborne radar to MIMO mode for environmental perception and establishes a sparse model between the array echo signal and the clutter scattering coefficient vector of the radar observation area. The estimated value of the clutter scattering coefficient of the observation area is obtained through sparse recovery technology. Then, the airborne radar is set to phased array mode for target detection. The estimated value of the clutter scattering coefficient is used to predict the clutter covariance matrix during target detection, and finally, a space-time two-dimensional adaptive filter is obtained to complete the non-uniform clutter suppression. This method can effectively suppress non-uniform clutter by acquiring prior clutter information. However, the drawback of this method is that the perception stage uses a convex optimization-type sparse recovery algorithm, which has a large computational load and high computational complexity, making it unsuitable for real-time clutter suppression.

[0005] Therefore, how to quickly and effectively suppress non-uniform clutter is one of the key technical problems faced by airborne radar. Summary of the Invention

[0006] To address the aforementioned problems in the existing technology, this invention provides a method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception. The technical problem to be solved by this invention is achieved through the following technical solution:

[0007] This invention provides a method for suppressing non-uniform clutter in airborne radar based on rapid environmental awareness, comprising the following steps:

[0008] When the airborne radar enters the sensing phase, several transmitting array elements transmit orthogonal signals, while the airborne radar acquires several first coherent pulses received by several receiving array elements.

[0009] Combining the orthogonal signals and the plurality of first coherent pulses, for the echo signal of each range unit, the number of discrete strong clutter blocks on each range loop of the airborne radar is estimated by beam scanning;

[0010] Establish a sparse recovery model of the echo signal and clutter block radar cross section for each range cell;

[0011] Based on the number of discrete strong clutter blocks, the sparse recovery model is solved using the orthogonal matching pursuit algorithm to obtain the clutter scattering coefficient estimate for each range ring;

[0012] The airborne radar enters the detection phase and transmits a linear frequency modulated signal, while simultaneously receiving several second coherent pulses;

[0013] Construct a structured clutter covariance matrix for each range cell based on the updated position of the airborne radar and the estimated clutter scattering coefficient.

[0014] The space-time two-dimensional adaptive weights are calculated using the structured clutter covariance matrix, and clutter suppression is performed using the space-time two-dimensional adaptive weights to obtain the clutter-suppressed signal.

[0015] In one embodiment of the present invention, combining the orthogonal signals and the plurality of first coherent pulses, for the echo signal of each range cell, beam scanning is used to estimate the number of discrete strong clutter blocks on each range loop of the airborne radar, including:

[0016] Using the airborne radar configuration parameters and platform flight parameters, beam scanning is performed on the echo signal of each range unit to obtain the first beam scan value;

[0017] Set a threshold value, and set the first beam scan value that is less than the threshold value to zero to obtain the second beam scan value;

[0018] A peak search is performed on the second beam scan value, and the number of peaks is taken as the number of discrete strong clutter blocks on each range ring.

[0019] In one embodiment of the present invention, the first beam scanning value is:

[0020]

[0021] in, This represents the first beam scan value. Indicates the scanning angle. The range of values ​​is And satisfy (·) H χ represents the conjugate transpose operation. l This represents the echo signal of the l-th distance cell. Represents the beam scanning weight vector.

[0022]

[0023] Indicates the Kronecker product. Represents the time-guided vector. Indicates the receiving guide vector. Represents the launch steering vector, a complex number. Used to eliminate the effect of transmit beam gain on beam scanning.

[0024]

[0025]

[0026]

[0027] j represents the imaginary unit, π represents pi, v0 represents the aircraft speed, sin represents the sine wave operation, λ represents the radar operating wavelength, and f r d represents the pulse repetition frequency. r =d represents the spacing between the receiving array elements, d t =N r d represents the distance between transmitting elements, d represents the distance between each element of the radar array, M1 represents the number of the first coherent pulses, and N r N represents the number of receiving array elements. t R(t) represents the number of transmitting array elements, and R(t) represents the correlation matrix of the transmitted signal. R(t) = ∫S(t)S H S(t)dt represents the orthogonal signal transmitted by the radar during the sensing phase.

[0028] In one embodiment of the present invention, peak search is performed on the second beam scan value, and the number of peaks is used as the number of discrete strong clutter blocks on each range ring, including:

[0029] The formula for peak search of the second beam scan value is:

[0030]

[0031] Among them, Ψ peak express A finite set consisting of the cone angles corresponding to all peaks in the equation. This represents the cone angle corresponding to the k-th clutter block in the l-th range cell. Indicates the scanning angle is The scan value at that time, Indicates the scanning angle is The scan value at that time, Indicates scan value The scan value at time, l represents the distance cell number, N c This indicates the number of clutter blocks on the l-th range ring;

[0032] The number of elements in the finite set of cone angles corresponding to all the peaks is taken as the number of peaks, and the number of discrete strong clutter blocks on each range ring is obtained.

[0033] In one embodiment of the present invention, the sparse recovery model is:

[0034] min||σ l ||0,st||χ l -H l σ l ||2≤ε

[0035] Where ||·||0 represents the 0-norm operation, ||·||2 represents the 2-norm operation, and χ l This represents the echo signal of the l-th distance cell. Let α be the vector representing the radar cross section of each clutter block in the l-th range cell. lk =ασ lk / r l 2 k = 1, ..., N c α represents a constant that is independent of clutter and depends only on radar system parameters, σ lk Let r represent the radar cross section of the k-th clutter block in the l-th range cell. l H represents the slant distance of the l-th distance unit. l M1N represents the distance unit l. t N r ×N c Wikipedia matrix, and

[0036]

[0037] Where α represents a constant that depends only on the radar system parameters, r lThis represents the slope distance of the l-th distance loop. M1N r N t ×1-dimensional spacetime steering vector Represents the time-guided vector. Indicates the receiving guide vector. R(t) represents the transmission steering vector, and R(t) represents the correlation matrix of the transmitted signal.

[0038] In one embodiment of the present invention, the sparse recovery model is solved using an orthogonal matching pursuit algorithm, taking into account the number of discrete strong clutter blocks, to obtain an estimate of the clutter scattering coefficient for each range ring, including:

[0039] The sparse recovery model is solved using the orthogonal matching pursuit algorithm, such that the number of iterations of the orthogonal matching pursuit algorithm is greater than or equal to the number of discrete strong clutter blocks, to obtain the clutter scattering coefficient estimate of each range ring.

[0040] In one embodiment of the present invention, constructing the structured clutter covariance matrix for each range loop based on the updated position of the airborne radar and the estimated clutter scattering coefficient includes:

[0041] Based on the first cone angle and first slant range of the clutter block relative to the airborne radar during the sensing phase, and the displacement of the airborne radar from the sensing phase to the detection phase, the second cone angle and second slant range of the clutter block relative to the airborne radar during the detection phase are calculated to obtain the updated position.

[0042] The structured clutter covariance matrix for each range ring is constructed based on the updated position and the clutter scattering coefficient estimate.

[0043] In one embodiment of the present invention, the structured clutter covariance matrix of each distance ring is:

[0044]

[0045]

[0046] in, Let |·| denote the structured clutter covariance matrix of the l-th distance cell, ∑· denotes the summation operation, and |·| 2 Let α represent the square of the modulus, where α is a constant independent of clutter blocks and only related to radar system parameters. This indicates the angle of the transmitted beam during the detection phase. Gain at that point, This represents the RCS of the k-th clutter block in the l-th range ring. Let I represent noise power, and let I represent the NM2-dimensional identity matrix. This represents the slant range of the l-th range loop to the radar during the detection phase. This represents the space-time steering vector of the k-th clutter block in the l-th range cell during the detection phase.

[0047] and Represented as:

[0048]

[0049]

[0050] Where N represents the number of array elements in the airborne radar, and M2 represents the number of second coherent pulses.

[0051] In one embodiment of the present invention, the calculation formula for the spatiotemporal two-dimensional adaptive weights is as follows:

[0052]

[0053] in(·) -1 This represents the inverse operation, v t Represents the desired spacetime steering vector. Let represent the structured clutter covariance matrix of the l-th distance loop.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0055] The method of this invention uses beam scanning to estimate the number of discrete strong clutter blocks on each range ring of the airborne radar, and combines the number of discrete strong clutter blocks in the orthogonal matching pursuit algorithm to solve the sparse recovery model of the sensing stage. This overcomes the problem that the stopping condition of the orthogonal matching pursuit algorithm cannot be determined when applied to environmental sensing, realizes rapid sensing of clutter environment, effectively reduces the algorithm complexity of environmental sensing process, and overcomes the problems of large amount of computation, high computational complexity and unfavorable to real-time clutter suppression in the prior art. Attached Figure Description

[0056] Figure 1 A flowchart illustrating a method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception, provided in an embodiment of the present invention.

[0057] Figure 2 A flowchart illustrating another method for suppressing non-uniform clutter in airborne radar based on rapid environmental awareness, provided in an embodiment of the present invention.

[0058] Figure 3 A schematic diagram illustrating the geometric relationship between an airborne radar and clutter blocks, provided as an embodiment of the present invention;

[0059] Figure 4 A schematic diagram of clutter block amplitude box line provided in an embodiment of the present invention;

[0060] Figure 5 A schematic diagram illustrating the geometric relationship between clutter blocks and radar when an airborne radar is moving, provided as an embodiment of the present invention;

[0061] Figure 6 This is a schematic diagram of a SAR image provided in an embodiment of the present invention;

[0062] Figure 7 This is a schematic diagram of a real clutter scene generated using SAR images, provided in an embodiment of the present invention.

[0063] Figure 8 A schematic diagram of a clutter scene reconstructed using an airborne radar non-uniform clutter suppression method based on rapid environmental perception, provided in an embodiment of the present invention.

[0064] Figure 9 A schematic diagram illustrating the results of space-time adaptive processing using the optimal processor, traditional environmental dynamic perception algorithm, airborne radar non-uniform clutter suppression method based on fast environmental perception, and SMI algorithm during the detection phase provided in this embodiment of the invention.

[0065] Figure 10 The diagram illustrates the code execution time of the traditional environmental dynamic perception algorithm and the airborne radar non-uniform clutter suppression method based on fast environmental perception provided in the embodiments of the present invention under different basis matrix dimensions. Detailed Implementation

[0066] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0067] Example 1

[0068] Please see Figure 1 and Figure 2 , Figure 1 This is a flowchart illustrating a method for suppressing non-uniform clutter in airborne radar based on rapid environmental awareness, provided in an embodiment of the present invention. Figure 2 This is a flowchart illustrating another airborne radar non-uniform clutter suppression method based on rapid environmental awareness, provided by an embodiment of the present invention. The clutter suppression method includes the following steps:

[0069] S1. The airborne radar enters the sensing phase, and several transmitting array elements transmit orthogonal signals. At the same time, the airborne radar acquires several first coherent pulses received by several receiving array elements.

[0070] Specifically, before the airborne radar detects a target, the system first enters the perception phase. At this time, the airborne radar operates in MIMO mode, transmitting orthogonal signals, N... tEach transmitter element emits orthogonal signals. N r The receiving array elements receive a total of M1 coherent pulses, of which N indicates that the antenna has N array elements. This indicates the floor function.

[0071] Please see Figure 3 , Figure 3 This is a schematic diagram illustrating the geometric relationship between an airborne radar and clutter blocks, provided as an embodiment of the present invention. Figure 3 In this diagram, the area enclosed by two concentric rings represents a range ring on the ground. Each range ring corresponds to a range cell. The shaded area on the range ring represents a clutter block. A single range ring contains multiple clutter blocks, and the clutter blocks on the same range ring have the same slant range as the radar. In this embodiment, the echo signal of the l-th range cell is represented as χ. l .

[0072] S2. Combining the orthogonal signals and the plurality of first coherent pulses, for the echo signal of each range unit, the number of discrete strong clutter blocks on each range loop of the airborne radar is estimated using beam scanning. Specifically, this includes the following steps:

[0073] S21. Using the airborne radar configuration parameters and platform flight parameters, perform beam scanning on the echo signal of each range unit to obtain the first beam scan value.

[0074] Specifically, using the airborne radar configuration parameters and platform flight parameters, beam scanning is performed on the echo signal of the l-th range cell. The beam scanning formula is:

[0075]

[0076] in, This represents the first beam scan value. Indicates the scanning angle. The range of values ​​is And satisfy (·) H χ represents the conjugate transpose operation. l This represents the echo signal of the l-th distance cell. Represents the beam scanning weight vector;

[0077]

[0078] in, Indicates the Kronecker product. Represents the time-oriented vector during the perception phase. This represents the receiving guidance vector during the perception phase. Represents the emission steering vector during the sensing phase, a complex number. Used to eliminate the influence of non-uniformity of transmit beam gain on beam scanning (in reality, the orthogonal signals transmitted by the radar cannot be completely orthogonal, so the transmit beam gain of the radar in the sensing phase is not strictly isotropic).

[0079]

[0080]

[0081]

[0082] Where j represents the imaginary unit, π represents pi, v0 represents the aircraft speed, sin represents the sinusoidal operation, λ represents the radar operating wavelength, and f r d represents the pulse repetition frequency. r =d represents the spacing between the receiving array elements, d t =N r d represents the distance between transmitting elements, d represents the distance between each element of the radar array, M1 represents the number of the first coherent pulses, and N r N represents the number of receiving array elements. t R(t) represents the number of transmitting array elements, and R(t) represents the correlation matrix of the transmitted signal. R(t) = ∫S(t)S H S(t)dt represents the orthogonal signal transmitted by the radar during the sensing phase.

[0083] S2. Set a threshold value, and set the first beam scan value that is less than the threshold value to zero to obtain the second beam scan value.

[0084] Specifically, in reality, due to the influence of sidelobe clutter, the beam scanning results... The curve is not smooth; random small peaks appear outside the angle corresponding to the discrete strong clutter block, which seriously affects the accuracy of subsequent estimation of the number of discrete strong clutter. To eliminate the influence of these random small peaks, this embodiment uses the following formula to set the first beam scan value less than the threshold to zero, obtaining the second beam scan value:

[0085]

[0086] in, Let ξ represent the threshold value, and let (0,1) represent a constant coefficient.

[0087] S3. Perform a peak search on the second beam scan value, and use the number of peaks as the number of discrete strong clutter blocks on each range ring. Specifically, this includes:

[0088] Specifically, the second beam scanning value is calculated using the following formula. Perform peak search:

[0089]

[0090] Among them, Ψ peak express A finite set consisting of the cone angles corresponding to all peaks in the equation. This represents the cone angle corresponding to the k-th clutter block in the l-th range cell. Indicates the scanning angle is The scan value at that time, Indicates the scanning angle is The scan value at that time, Indicates the scanning angle is The scan value at time, l represents the distance ring number, N c This represents the number of clutter blocks on the l-th range ring. Set Ψ peak The number of elements in the peak value is the number of peak values.

[0091] Then, the finite set Ψ consisting of the cone angles corresponding to all peaks. peak The number of elements in the peak value is used as the number of peak values. Number of peaks This is the number of discrete strong clutter blocks on the l-th range ring.

[0092] Please see Figure 4 , Figure 4 This is a schematic diagram of the clutter block amplitude box curve provided in an embodiment of the present invention. In this embodiment, the upper edge of the clutter block amplitude box curve is used as the threshold. Clutter blocks with an RCS greater than the threshold are strong clutter blocks, and clutter blocks with an RCS less than or equal to the threshold are weak clutter blocks.

[0093] S3. Establish a sparse recovery model of the echo signal and clutter block radar cross section for each range cell.

[0094] Specifically, the sparse recovery model of the echo signal and the radar cross section (RCS) of the clutter block in the sensing stage of the l-th range cell is established as follows:

[0095] min||σ l ||0,st||χ l -H l σ l ||2≤ε

[0096] Where ||·||0 represents the 0-norm operation, ||·||2 represents the 2-norm operation, and χ l This represents the echo signal of the l-th distance cell. Let α be the vector representing the radar cross section (RCS) of each clutter block in the l-th range cell. lk =ασ lk / r l 2 k = 1, ..., N c α represents a constant that is independent of clutter and depends only on radar system parameters, σ lk Let r represent the radar cross section of the k-th clutter block in the l-th range cell. l H represents the slant distance of the l-th distance unit. l M1N represents the distance unit l. t N r ×N c The vidiquity matrix (the dictionary matrix corresponding to the l-th distance ring), and

[0097]

[0098] Where α represents a constant that depends only on the radar system parameters, r l This represents the slant range from the l-th range loop to the radar. M1N r N t ×1-dimensional spacetime steering vector Represents the time-guided vector. Indicates the receiving guide vector. R(t) represents the transmission steering vector, and R(t) represents the correlation matrix of the transmitted signal.

[0099] S4. Combining the number of discrete strong clutter blocks, the sparse recovery model is solved using the orthogonal matching pursuit algorithm to obtain the estimated clutter scattering coefficient for each range ring. Specific steps include:

[0100] Specifically, the sparse recovery model is solved using the orthogonal matching pursuit algorithm, such that the number of iterations of the orthogonal matching pursuit algorithm is greater than or equal to the number of discrete strong clutter blocks, to obtain the clutter scattering coefficient estimate for each range ring.

[0101] The specific steps for solving the sparse recovery model using the orthogonal matching algorithm are as follows:

[0102] (1) Initialize residual r0 = χ l RCS estimation vector Atomic support set Atomic number set Iteration number t = 1;

[0103] (2) Calculate the residual r t M1N with the l-th distance unit t N r ×N c Wikipedia H lThe inner product of each column vector (called an atom) is calculated, and the atom with the largest inner product value is selected and recorded as atom number i. t );

[0104] (3) Add the atoms selected in step (2) to the atom support set:

[0105]

[0106] in, This represents the atom with the largest inner product value with the residual. Indicates the i-th (t) The cone angles corresponding to each column vector;

[0107] Atomic numbers are added to the set of atomic numbers:

[0108] Λ t =Λ t-1 ∪{i (t)};

[0109] (4) Calculate the RCS vector σ using the least squares estimation method of the following formula. l Approximate estimate:

[0110]

[0111] Where, χ l This represents the echo signal of the l-th distance unit;

[0112] (5) Update residuals

[0113]

[0114] (6) Determine whether t is less than the number of discrete strong clutter blocks on the l-th distance ring. If yes, return to step (2) to continue iteration; if no, stop iteration and output the estimated clutter scattering coefficient. That is, the RCS vector estimation result is

[0115] S5. The airborne radar enters the detection phase and transmits a linear frequency modulated signal, while simultaneously receiving several second coherent pulses.

[0116] Specifically, after completing clutter environment perception, the airborne radar enters the detection phase. At this time, the radar operates in phased array mode, with N transmitting array elements transmitting conventional linear frequency modulated signals for target detection, and N receiving array elements receiving M2 second coherent pulses. The remaining system parameters (radar operating wavelength, transmitted signal bandwidth, pulse repetition frequency, etc.) are the same as in the perception phase.

[0117] S6. Construct the structured clutter covariance matrix for each range loop based on the updated position of the airborne radar and the estimated clutter scattering coefficient. Specifically, this includes the following steps:

[0118] S61. Based on the first cone angle and first slant range of the clutter block relative to the airborne radar during the sensing phase and the displacement of the airborne radar from the sensing phase to the detection phase, calculate the second cone angle and second slant range of the clutter block relative to the airborne radar during the detection phase to obtain the updated position.

[0119] Specifically, after the carrier platform moves to a new position, the slant range of the clutter block to the radar also changes with the carrier platform's position. Therefore, before calculating the covariance matrix using the RCS estimation results from the sensing phase, it is necessary to update the clutter block's position information. For example... Figure 5 As shown, Figure 5 This is a schematic diagram illustrating the geometric relationship between clutter and radar when an airborne radar is moving, provided as an embodiment of the present invention. It is assumed that the airborne radar is located at point A during the sensing phase, and the first cone angle and first slant range of the clutter relative to the radar are respectively... And r, the detection phase is located in At this point, the clutter block's second cone angle and second slant range relative to the radar are respectively... and The displacement of the airborne radar from the sensing phase to the detection phase is Δy. Based on trigonometric functions, the update formulas for the second cone angle and the second slant range can be obtained as follows:

[0120]

[0121]

[0122] in, Let r be the first cone angle, r be the first slant range, and Δy be the displacement of the airborne radar from the sensing phase to the detection phase.

[0123] S62. Construct the structured clutter covariance matrix for each range ring based on the updated position and the estimated clutter scattering coefficient.

[0124] Specifically, after obtaining the updated position and clutter scattering coefficient estimates from the airborne radar, the prior covariance matrix is ​​constructed according to the following formula. The structured clutter covariance matrix of the l-th range loop can be expressed as:

[0125]

[0126]

[0127] in, Let |·| denote the structured clutter covariance matrix of the l-th distance cell, ∑· denotes the summation operation, and |·| 2 Let α represent the square of the modulus, where α is a constant independent of clutter blocks and only related to radar system parameters. This indicates the angle of the transmitted beam during the detection phase. Gain at that point, This represents the RCS of the k-th clutter block in the l-th range ring. Let I represent noise power, and let I represent the NM2-dimensional identity matrix. This represents the slant range of the l-th range loop to the radar during the detection phase. This represents the space-time steering vector of the k-th clutter block in the l-th range cell during the detection phase.

[0128] and Represented as:

[0129]

[0130]

[0131] Where N represents the number of array elements in the airborne radar, and M2 represents the number of second coherent pulses.

[0132] S7. Calculate the space-time two-dimensional adaptive weights using the structured clutter covariance matrix, and use the space-time two-dimensional adaptive weights to suppress clutter, thereby obtaining the clutter-suppressed signal.

[0133] Specifically, after obtaining the structured clutter covariance matrix of the l-th distance loop, the spatiotemporal two-dimensional adaptive weights are calculated as follows:

[0134]

[0135] in(·) -1 This represents the inverse operation, v t Represents the desired spacetime steering vector. Let represent the structured clutter covariance matrix of the l-th distance loop.

[0136] Then, the space-time two-dimensional adaptive weights are multiplied with the echo signal of the l-th distance cell to obtain the clutter-suppressed signal.

[0137] In this embodiment, beam scanning estimates the number of discrete strong clutter blocks on each range ring of the airborne radar, and combines the number of discrete strong clutter blocks in the orthogonal matched pursuit algorithm to solve the sparse recovery model in the sensing stage. The number of discrete strong clutter blocks is used as the iteration stopping condition of the orthogonal matched pursuit algorithm, which overcomes the problem that the stopping condition of the orthogonal matched pursuit algorithm cannot be determined when applied to environmental perception.

[0138] This embodiment utilizes the orthogonal matching pursuit algorithm to solve the sparse recovery problem, achieving rapid perception of clutter environment. It effectively reduces the algorithmic complexity of the environment perception process and overcomes the problems of large computational load, high computational complexity, and unfavorable conditions for real-time clutter suppression in existing technologies.

[0139] Furthermore, this embodiment illustrates the method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception through simulation experiments.

[0140] 1. Simulation experimental conditions:

[0141] The hardware platform for the simulation experiment in this embodiment is: Intel(R) Core(TM) i7-10750H CPU with a main frequency of 2.60GHz and 32GB of memory.

[0142] The software platform for the simulation experiment in this embodiment is: Windows 10 operating system and MATLAB R2020a.

[0143] The simulation parameters for this embodiment are set as follows: the airborne front-side looking array radar uses a 1×100 half-wavelength equidistant linear array with an element spacing of 12cm, a radar operating wavelength of 24cm, a transmit signal bandwidth of 2MHz, and a pulse repetition frequency of 2500Hz. The carrier platform's flight speed is 140m / s, and the carrier altitude is 5000m. In the sensing phase, starting with the first element, elements are selected at intervals of 10 to form the transmitting array. Elements 1 to 10 form the receiving array, receiving 4 coherent pulses. In the detection phase, the radar receives 16 coherent pulses, and the remaining operating parameters are the same as in the sensing phase.

[0144] 2. Simulation content and result analysis:

[0145] This embodiment uses a SAR image as the actual clutter scene, such as Figure 6 As shown, Figure 6 This is a schematic diagram of a SAR image provided in an embodiment of the present invention. Each pixel in the image serves as a clutter scattering source, and the pixel value represents the RCS of the scattering source. Figure 7 As shown, Figure 7 This is a schematic diagram of a real clutter scene generated using SAR images, provided in an embodiment of the present invention. In the experiment, the actual clutter scene was divided into clutter blocks at intervals of 0.25°, and the sum of the pixel values ​​of the SAR image pixels within the clutter block was used as the RCS of the clutter block.

[0146] Under the above experimental conditions, the noise-to-clutter ratio of the echo signal during the sensing phase was set to 25dB, and the grid spacing was 0.25° when constructing the dictionary matrix. For example... Figure 8 As shown, Figure 8This is a schematic diagram of a clutter scene reconstructed using an airborne radar non-uniform clutter suppression method based on rapid environmental perception, provided in an embodiment of the present invention. Figure 7 Compared with actual clutter scenarios, the method in this embodiment can effectively reconstruct discrete strong clutter blocks, which are the main influencing factors for target detection.

[0147] During the detection phase, the echo clutter ratio was set to 40dB. This embodiment employs an optimal processor, a traditional dynamic environment sensing algorithm, an airborne radar non-uniform clutter suppression method based on fast environment sensing, and a sampling covariance matrix inversion (SMI) algorithm for clutter suppression. This embodiment simulates the improvement factor, defined as the ratio of the output SNR to the input SNR, used to measure the clutter suppression capability of the algorithm. After clutter suppression using this embodiment, in [-f r / 2,-f r Within the range of / 2], multiple frequency points are selected, and the values ​​of the improvement factor corresponding to each frequency point are connected to obtain the improvement factor curve. The result is as follows: Figure 9 As shown, Figure 9 This diagram illustrates the results of space-time adaptive processing during the detection phase provided in this embodiment of the invention, utilizing an optimal processor, a traditional dynamic environmental perception algorithm, an airborne radar non-uniform clutter suppression method based on fast environmental perception, and the SMI algorithm. The following is a related illustration... Figure 9 The simulation results further illustrate the effects of the present invention.

[0148] Specifically, Figure 9 The curve illustrates the change in the improvement factor with normalized Doppler frequency. The horizontal axis represents the normalized Doppler frequency, and the vertical axis represents the improvement factor, with the unit being decibels (dB). Figure 9 As shown in the simulation results, the traditional SMI algorithm cannot accurately estimate the clutter covariance matrix due to its inability to obtain sufficient independent and identically distributed training samples, resulting in poor clutter suppression performance. The algorithm proposed in this embodiment can effectively predict the clutter covariance matrix, thus significantly improving clutter suppression performance in non-uniform environments. Compared with traditional environmental dynamic sensing algorithms, the algorithm proposed in this embodiment achieves similar performance to traditional environmental dynamic sensing algorithms in different Doppler channels and is close to the performance of the optimal processor.

[0149] This embodiment compares the runtime of traditional environmental dynamic perception algorithms with that of the proposed algorithm to illustrate the computational complexity advantage of the proposed algorithm. The runtime of the code is calculated using the commonly used MATLAB commands `tic` and `toc`. To better compare the runtime of the two algorithms, this embodiment sets the number of coherent pulses to a fixed 4, and the number of transmitting and receiving array elements to be the same. The experiment simulates a scenario with N transmitting array elements. tThe running times of the two algorithms when the time is 4, 6, 8, 10, 12, for example... Figure 10 As shown, Figure 10 The diagram illustrates the code execution time of the traditional environmental dynamic perception algorithm and the airborne radar non-uniform clutter suppression method based on fast environmental perception provided in the embodiments of the present invention under different basis matrix dimensions.

[0150] Depend on Figure 10 Simulation results show that, for different numbers of matrix elements (dictionary matrices of different dimensions), the code execution time of the algorithm proposed in this embodiment is much shorter than that of traditional environment dynamic perception algorithms. Therefore, the algorithm in this embodiment has lower computational complexity compared to traditional environment dynamic perception algorithms.

[0151] The simulation results above verify the correctness, effectiveness, and reliability of the non-uniform clutter suppression method for airborne radar based on rapid environmental awareness in this embodiment.

[0152] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception, characterized in that, Including the following steps: When the airborne radar enters the sensing phase, several transmitting array elements transmit orthogonal signals, while the airborne radar acquires several first coherent pulses received by several receiving array elements. Combining the orthogonal signals and the plurality of first coherent pulses, for the echo signal of each range cell, the number of discrete strong clutter blocks on each range loop of the airborne radar is estimated by beam scanning; Establish a sparse recovery model of the echo signal and clutter block radar cross section for each range cell; Based on the number of discrete strong clutter blocks, the sparse recovery model is solved using the orthogonal matching pursuit algorithm. When it is determined that the number of iterations of the orthogonal matching pursuit algorithm is not less than the number of discrete strong clutter blocks, the clutter scattering coefficient estimate of each range ring is obtained. The airborne radar enters the detection phase and transmits a linear frequency modulated signal, while simultaneously receiving several second coherent pulses; Construct a structured clutter covariance matrix for each range cell based on the updated position of the airborne radar and the estimated clutter scattering coefficient. The space-time two-dimensional adaptive weights are calculated using the structured clutter covariance matrix, and clutter suppression is performed using the space-time two-dimensional adaptive weights to obtain the clutter-suppressed signal.

2. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 1, characterized in that, Combining the orthogonal signals and the plurality of first coherent pulses, for the echo signal of each range cell, the number of discrete strong clutter blocks on each range loop of the airborne radar is estimated using beam scanning, including: Using the airborne radar configuration parameters and platform flight parameters, beam scanning is performed on the echo signal of each range unit to obtain the first beam scan value; Set a threshold value, and set the first beam scan value that is less than the threshold value to zero to obtain the second beam scan value; A peak search is performed on the second beam scan value, and the number of peaks is taken as the number of discrete strong clutter blocks on each range ring.

3. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 2, characterized in that, The first beam scanning value is: in, This represents the first beam scan value. Indicates the scanning angle. The range of values ​​is And satisfy φ represents the scanning angle. The projection angle on the horizontal plane, i.e., the azimuth angle, θ represents the angle between the scanning angle and the horizontal plane, i.e., the elevation angle, (·). H χ represents the conjugate transpose operation. l This represents the echo signal of the l-th distance cell. Represents the beam scanning weight vector. Indicates the Kronecker product. Represents the time-guided vector. Indicates the receiving guide vector. Represents the launch steering vector, a complex number. Used to eliminate the effect of transmit beam gain on beam scanning. j represents the imaginary unit, π represents pi, v0 represents the aircraft speed, sin represents the sine wave operation, λ represents the radar operating wavelength, and f r d represents the pulse repetition frequency. r =d represents the spacing between the receiving array elements, d t =N r d represents the distance between transmitting elements, d represents the distance between each element of the radar array, M1 represents the number of the first coherent pulses, and N r N represents the number of receiving array elements. t R(t) represents the number of transmitting array elements, and R(t) represents the correlation matrix of the transmitted signal. R(t) = ∫S(t)S H S(t)dt represents the orthogonal signal transmitted by the radar during the sensing phase.

4. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 2, characterized in that, A peak search is performed on the second beam scan value, and the number of peaks is used as the number of discrete strong clutter blocks on each range ring, including: The formula for peak search of the second beam scan value is: Among them, Ψ peak express A finite set consisting of the cone angles corresponding to all peaks in the equation. This represents the cone angle corresponding to the k-th clutter block in the l-th range cell. Indicates the scanning angle is The scan value at that time, Indicates the scanning angle is The scan value at that time, Indicates the scanning angle is The scan value at time, l represents the distance ring number, N c This indicates the number of clutter blocks on the l-th range ring; The number of elements in the finite set of cone angles corresponding to all the peaks is taken as the number of peaks, and the number of discrete strong clutter blocks on each range ring is obtained.

5. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 1, characterized in that, The sparse recovery model is as follows: min||σ l ||0,st||x l -H l s l ||2≤e Where ||·||0 represents the 0-norm operation, ||·||2 represents the 2-norm operation, and χ l This represents the echo signal of the l-th distance cell. Let represent the vector formed by the radar cross-sections of each clutter block in the l-th range cell. α represents a constant that is independent of clutter and depends only on radar system parameters, σ lk Let r represent the radar cross section of the k-th clutter block in the l-th range cell. l H represents the slant distance of the l-th distance unit. l M1N represents the distance unit l. t N r ×N c Wikipedia matrix, and Where, r l This represents the slope distance of the l-th distance loop. M1N r N t ×1-dimensional spacetime steering vector Represents the time-guided vector. Indicates the receiving guide vector. R(t) represents the transmission steering vector, and R(t) represents the correlation matrix of the transmitted signal.

6. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 1, characterized in that, Based on the updated position of the airborne radar and the estimated clutter scattering coefficient, a structured clutter covariance matrix is ​​constructed for each range cell, including: Based on the first cone angle and first slant range of the clutter block relative to the airborne radar during the sensing phase, and the displacement of the airborne radar from the sensing phase to the detection phase, the second cone angle and second slant range of the clutter block relative to the airborne radar during the detection phase are calculated to obtain the updated position. A structured clutter covariance matrix is ​​constructed for each range cell based on the updated position and the clutter scattering coefficient estimate.

7. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 1, characterized in that, The structured noise covariance matrix is: in, Let |·| represent the structured clutter covariance matrix of the l-th distance cell, ∑· represent the summation operation, and |·| 2 Let α represent the square of the modulus, where α is a constant independent of clutter and only related to the radar system parameters. This indicates the angle of the transmitted beam during the detection phase. Gain at that point, This represents the RCS of the k-th clutter block in the l-th range ring. Let I represent noise power, and let I represent the NM2-dimensional identity matrix. This represents the slant range of the l-th range loop to the radar during the detection phase. This represents the space-time steering vector of the k-th clutter block in the l-th range cell during the detection phase. and Represented as: Where N represents the number of array elements in the airborne radar, and M2 represents the number of second coherent pulses.

8. The method for suppressing non-uniform clutter in airborne radar based on rapid environmental perception according to claim 1, characterized in that, The formula for calculating the space-time two-dimensional adaptive weights is as follows: in(·) -1 This indicates the inverse operation, v t Represents the desired spacetime steering vector. Let represent the structured clutter covariance matrix of the l-th distance loop.

Images